An HDP-HMM for systems with state persistence

By Emily B. Fox, Erik B. Sudderth, Michael I. Jordan, Alan S. Willsky. International Conference on Machine Learning, 2008

330 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Signal Processing

View Paper: https://www.semanticscholar.org/paper/28627cecfd911455ffabfaf7acf248f7d5b78d1b

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Signal Processing

Citations in the corpus

Bayesian nonparametric hidden semi-Markov models (2012, methodology): “The weak-limit sampler for an HDP-HMM (Fox et al., 2008) constructs a finite approximation to the HDP transitions prior with finite L-dimensional Dirichlet distributions, motivated by the fact that the infinite limit of such a construction converges in distribution to a true HDP: β|γ ∼ Dir(γ/L, . .…”
A Non-parametric Learning Method for Confidently Estimating Patient's Clinical State and Dynamics (2016, methodology): “Gibbs sampler [10], Beam sampler [25], variational Bayes [6, 7]) can be utilized to estimate the parameters of the patient’s stochastic model (4) given the data {yt}t∈T .”
A Sticky HDP-HMM With Application to Speaker Diarization (2009, methodology): “A first attempt at developing such models is presented in Fox et al. (2009). An inspiration for the sticky HDP-HMM actually came from considering the original HDP-HMM as a prior for an SLDS.”
Bayesian Nonparametric Inference of Switching Dynamic Linear Models (2010, methodology): “Also, we note that step (2) involves a fairly straightforw a d extension of the sampling method developed in [14] for the simpler HDP-HMM model; the other st eps, however, involve new constructs, as they involve capturing and dealing with the temporal dynamics of the underlying continuous state mode”
Online Variational Inference for the Hierarchical Dirichlet Process (2011, methodology): “We will begin by reviewing Sethuraman’s stick-breaking construction of the HDP [14].”
Bayesian Nonparametrics: Hierarchical Bayesian nonparametric models with applications (2010, methodology): “(25) Fox et al. (2009) further place a vague gamma prior on α + κ and a beta prior on κ/(α + κ).”
Speaker Diarization: A Review of Recent Research (2010, unknown): “The HDP-HMM has shown promise in diarization [129], yielding similar performance to the standard agglomerative HMM with GMM emissions, while requiring very little hyperparameter tuning and providing a statistically sound model.”
Statistical Model Aggregation via Parameter Matching (2019, unknown): “We compare against local HDP-HMMs as well as two strong competitors, an identical sticky HDP-HMM model but trained on all six sequences jointly, and an alternate matching scheme based on k -means clustering that clusters together states discovered by the individual HDP-HMMs.”
Sharing Features among Dynamical Systems with Beta Processes (2009, methodology): “Other recent approaches to Bayesian nonparametric represe ntations of time series include the HDPHMM [2, 4, 5, 21] and the infinite factorial HMM [24].”
Learning grounded finite-state representations from unstructured demonstrations (2015, methodology): “The HDP-HMM allows for a potentially infinite number of latent modes to be discovered in a fully Bayesian manner, reducing the risk of overfitting and the need for prior knowledge .”
Stick-breaking autoregressive processes (2011, methodology): “Fox et al. (2008) develop a hidden Markov model with an infinite number of states and apply their method to regime-switching in financial time series.”